> That looks like the one. > Since we can generate the discards on one AP and one SM, we aren't using > that as an indication that an AP is overloaded. > I assume inbound packets destined for the SM are dropped at the AP due to > MIR settings, thus creating a discard.
# snmpget -v2c -c Canopyro -m ALL 169.254.1.1 .1.3.6.1.4.1.161.19.3.1.7.35.0 SNMPv2-SMI::enterprises.161.19.3.1.7.35.0 = No Such Instance currently exists at this OID Bummer. What am I doing wrong? >>> The only existing tool built into an AP that does this would be under >>> Statistics > Throughput and enabling Throughput Monitoring. >>> This will tell you when you are tossing xx% of RF packets due to an >>> "overload" condition. >>> We graph 'RF out discards' on each AP and connection but we seem to get >>> them >> >> Is this what you graph? >> >> >> http://support.cambiumnetworks.com/framed/onlinetools/content.WHISP-APS-MIB.html#rfOutDiscardRate >> >> Apparently its not telling you much? >> >> >>> even during low utilization periods. >>> With 450, we can generate RF discards in the lab with a single AP and >>> single >>> SM so we aren't really sure when an AP is overloaded. >>> >>> Please head over to Cambium's community forum and give Kudos to these >>> posts: >>> >>> http://community.cambiumnetworks.com/t5/Your-Ideas/Utilization-per-Sector/idi-p/37968 >>> >>> http://community.cambiumnetworks.com/t5/Your-Ideas/450-AP-Throughput-capacity/idi-p/36808 >>> >>> Alternatively, you can input your AP's session info (current up and down >>> modulations) into the Capacity Planner available here: >>> https://support.cambiumnetworks.com/files/pmp450/ >>> That should give you a theoretical max available bandwidth. >>> >>> >>> On 1/26/2015 11:57 AM, Matt wrote: >>>> >>>> We graph all the ethernet Interfaces on our FSK Canopy AP's etc with >>>> MRTG. Is there a way to graph the actual utilization to see when an >>>> AP is maxed? I know an AP with all 2X clients should have 10mbps >>>> downstream which most ours do. What I cant tell is when an AP has a >>>> mix of 1X and 2X if the AP is maxed out or not by the graph. Same >>>> principle for 450 AP's. >>>> >>>> Perhaps there is a packet queue on the AP that gets congested at times >>>> like this I can graph on five minute intervals? Something like that? >>> >>> >
